Technology company Google has announced the introduction of two new models of artificial intelligence, ‘Gemini 4 Carbon’ and ‘Genie 4’.

The company had introduced new AI models just two weeks ago. In the latest development, one model has been developed for cybersecurity tasks, while the other focuses on the ability to create and operate digital worlds.

According to Google, Gemini 4 Carbon is its most advanced AI model to date. According to the company, it performed better than some other models, including Gemini 4 Argon and OpenAI’s GPT 6 Astra, in benchmark tests. This model is specifically designed to help cybersecurity experts identify and fix vulnerabilities in computer systems.

On the other hand, Genie 4 is a ‘world model’, which aims to create digital worlds based on simulations of real or imaginary environments. While typical chatbots predict the next words to be spoken in response to a user’s query, world models attempt to model an environment, its elements, and their interactions. Such technology could be useful in a variety of fields, including robotics, virtual environments, and video game development.

Gemini 4 also includes audio support. This will allow developers to add background sounds and dialogue to their digital worlds, helping to make these environments more realistic and dynamic.

Neither model is currently available to the general public. According to Google, access to Gemini 4 Carbon is being given to select cybersecurity experts. No further details have been released about the widespread availability of these models to general users or developers.